Liver Function Tests in Patients P. J. CLARKE, From the
Receiving Parenteral Nutrition
F.R.C.S., M. J. BALL, M.R.C.P., M.D.,
AND
Nuffield Departments of Surgery and Clinical Biochemistry,
ABSTRACT. A 5-year prospective study
was
performed
M. G. W. KETTLEWELL, F.R.C.S., M.CHIR. John
Radcliffe Hospital, Oxford, United Kingdom
TPN, but overall abnormal LFTs did not appear more common in these patients than in those without obvious sepsis. Patients with malignant disease, those requiring long-term TPN, and those requiring a nonstandard TPN regimen were more likely to develop raised LFTs. ( Journal of Parenteral and Enteral Nutrition 15: 54-59, 1991)
to
monitor liver function tests (LFTs) in patients receiving total parenteral nutrition (TPN). A gradual and progressive rise was seen in the plasma concentration of bilirubin, aspartate transaminase, and alkaline phosphatase. The rate of rise was not increased in patients with LFT abnormalities before the start of TPN. Half of the patients had an episode of sepsis during
There are several complications which may arise durtotal parenteral nutrition (TPN). Some are related to the insertion and care of a central venous line, and others are biochemical and physiologic disorders. The latter are often diagnosed because they cause abnormal results on biochemical or hematologic analysis. Some of these derangements are trivial and corrected by modifying the prescription of the TPN, but derangement of liver function tests (LFTs) are common, 1,2 often persistent, and may cause clinical concern. Abnormal LFTs occur in as many as two thirds of patients fed intravenously for 2 weeks or more.’ The cause and significance of the abnormalities is unclear. In the majority of patients the changes are not extreme, and values return to within the normal range on the cessation of TPN and reintroduction of enteral nutrition. However, some patients may develop progressive hepatic impairment which can prove fatal, and there appears no obvious way of predicting which patients will be affected. The aim of this study was to monitor LFTs prospectively in a large group of patients to determine the incidence of abnormalities, to ascertain their nature, and to investigate whether any particular factors may be used to predict those patients at greatest risk of developing deranged LFTs and liver disease.
Most patients (75%) were fed parenterally using a daily regimen containing 14 g of nitrogen as amino acids (Synthamin, Travenol, or Aminoplasmol, Braun, FRG), 250 g of glucose, and 500 ml of a 20% triglyceride emulsion (Intralipid, KabiVitrum, or Lipofundin S, Braun).
ing
The remaining patients received an infusion which was tailored to their individual requirements of calories and nitrogen, because the standard feed was inappropriate due to fluid overload, renal impairment, or particular biochemical abnormalities. Blood samples were taken before the start of TPN. Hemoglobin, white cell, and platelet count were measured on a Coulter S-plus analyzer. Clotting time was measured and plasma electrolytes, urea, and LFTs (albumin, aspartate transaminase [AST], bilirubin, and alkaline phosphatase) were measured on a monitor parallel analyzer using the company’s recommended protocols. The basic electrolyte and vitamin content of the TPN was as shown on Table I but was adjusted according to the patients’ plasma electrolyte results and fluid balance. Patients received blood transfusion when clinically warranted. LFTs were repeated at least once each week while the patient received TPN and an &dquo;abnormal&dquo; value for any parameter was taken as one greater than twice the upper limit of the standard reference or &dquo;normal&dquo; range for the laboratory, which is based on the mean ± 2 SD for a healthy population. This was chosen because the reference range in sick patients is not the same as that derived from &dquo;healthy&dquo; people and small differences are often expected. Values of at least twice the healthy range would seem more likely to warrant consideration, and for our laboratory this was a plasma bilirubin level of >34 j.Lmoljliter, an AST >70 IU/liter, and an alkaline phosphatase concentration >600 IU/liter. Plasma albumin was not considered in the analysis because levels tend to be more affected by fluid balance and catabolism than hepatic synthesis. If LFT abnormalities developed during feeding, specific causes such as sepsis or a reaction to drug administration were considered. The feeding regimen was discontinued only if the abnormalities progressed very rap-
PATIENTS AND METHODS
All adults receiving TPN in Oxford during the 5-year period between 1980 and 1985 were studied prospectively. Records were made of the clinical details of each patient, the feeding regimen(s) used, any episodes of sepsis, and the outcome of feeding. The diagnosis of sepsis was made on the basis of fever, raised white cell count, known or suspected source of infection, and in many cases a positive microbiologic culture. Patients with known, antecedent chronic liver disease were excluded from the analysis.
Reprint requests: Miss P. J. Clarke, Nuffield Department of Surgery, John Radcliffe Hospital, Oxford OX3 9DU, UK. 54
Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on March 30, 2015
55
idly and discontinuation
was considered clinicallv showed a gradual but progressive increase in alkaline analyzed for the group as a phosphatase and AST levels and a small increase in whole. In addition, the data were analyzed for a number plasma bilirubin, and after 4 weeks of feeding the biliof subgroups and a comparison made between the follow- rubin was abnormal in 31%, the AST in 27%. and the ing : (1) patients whose pre-TPN LFTs were normal and alkaline phosphatase in 32%. those whose results were abnormal; (2) patients who had episodes of sepsis during the course of parenteral feeding Effect of Episodes of Sepsis During TPN and those who did not; (3) patients who survived to leave the study, 208 patients (49.5%) had an episode hospital and those who died during, or soon after TPN of During LFTs in patients without sepsis are shown in sepsis. stopped; and (4) patients who were fed for more than 28 3 and show a similar pattern to that in the total Figure days were considered as a separate group. Half of all the patients without sepsis Finally, an additional analysis was performed on the group (Fig. 1). abnormal LFTs. The percentages eventually developed 308 patients with normal pre-TPN LFTs, who received of septic and nonseptic patients with abnormal LFTs at A TPN during the last 3 years of the study. detailed each week are shown in Figure 4. The trend was for study was made of the 64 patients in this group whose of LFTs LFTs became abnormal during feeding in order to ascer- septic patients to have greater derangements but the only statistically signiftain whether there was any common factor which pre- than nonseptic patients, icant difference between the two was a higher disposed to this problem. Statistical analyses of the data bilirubin level for weeks 2 and 3 ingroups septic patients (p were performed using the Mann-Whitney U-test x2 anal4
Receiving Parenteral Nutrition
F.R.C.S., M. J. BALL, M.R.C.P., M.D.,
AND
Nuffield Departments of Surgery and Clinical Biochemistry,
ABSTRACT. A 5-year prospective study
was
performed
M. G. W. KETTLEWELL, F.R.C.S., M.CHIR. John
Radcliffe Hospital, Oxford, United Kingdom
TPN, but overall abnormal LFTs did not appear more common in these patients than in those without obvious sepsis. Patients with malignant disease, those requiring long-term TPN, and those requiring a nonstandard TPN regimen were more likely to develop raised LFTs. ( Journal of Parenteral and Enteral Nutrition 15: 54-59, 1991)
to
monitor liver function tests (LFTs) in patients receiving total parenteral nutrition (TPN). A gradual and progressive rise was seen in the plasma concentration of bilirubin, aspartate transaminase, and alkaline phosphatase. The rate of rise was not increased in patients with LFT abnormalities before the start of TPN. Half of the patients had an episode of sepsis during
There are several complications which may arise durtotal parenteral nutrition (TPN). Some are related to the insertion and care of a central venous line, and others are biochemical and physiologic disorders. The latter are often diagnosed because they cause abnormal results on biochemical or hematologic analysis. Some of these derangements are trivial and corrected by modifying the prescription of the TPN, but derangement of liver function tests (LFTs) are common, 1,2 often persistent, and may cause clinical concern. Abnormal LFTs occur in as many as two thirds of patients fed intravenously for 2 weeks or more.’ The cause and significance of the abnormalities is unclear. In the majority of patients the changes are not extreme, and values return to within the normal range on the cessation of TPN and reintroduction of enteral nutrition. However, some patients may develop progressive hepatic impairment which can prove fatal, and there appears no obvious way of predicting which patients will be affected. The aim of this study was to monitor LFTs prospectively in a large group of patients to determine the incidence of abnormalities, to ascertain their nature, and to investigate whether any particular factors may be used to predict those patients at greatest risk of developing deranged LFTs and liver disease.
Most patients (75%) were fed parenterally using a daily regimen containing 14 g of nitrogen as amino acids (Synthamin, Travenol, or Aminoplasmol, Braun, FRG), 250 g of glucose, and 500 ml of a 20% triglyceride emulsion (Intralipid, KabiVitrum, or Lipofundin S, Braun).
ing
The remaining patients received an infusion which was tailored to their individual requirements of calories and nitrogen, because the standard feed was inappropriate due to fluid overload, renal impairment, or particular biochemical abnormalities. Blood samples were taken before the start of TPN. Hemoglobin, white cell, and platelet count were measured on a Coulter S-plus analyzer. Clotting time was measured and plasma electrolytes, urea, and LFTs (albumin, aspartate transaminase [AST], bilirubin, and alkaline phosphatase) were measured on a monitor parallel analyzer using the company’s recommended protocols. The basic electrolyte and vitamin content of the TPN was as shown on Table I but was adjusted according to the patients’ plasma electrolyte results and fluid balance. Patients received blood transfusion when clinically warranted. LFTs were repeated at least once each week while the patient received TPN and an &dquo;abnormal&dquo; value for any parameter was taken as one greater than twice the upper limit of the standard reference or &dquo;normal&dquo; range for the laboratory, which is based on the mean ± 2 SD for a healthy population. This was chosen because the reference range in sick patients is not the same as that derived from &dquo;healthy&dquo; people and small differences are often expected. Values of at least twice the healthy range would seem more likely to warrant consideration, and for our laboratory this was a plasma bilirubin level of >34 j.Lmoljliter, an AST >70 IU/liter, and an alkaline phosphatase concentration >600 IU/liter. Plasma albumin was not considered in the analysis because levels tend to be more affected by fluid balance and catabolism than hepatic synthesis. If LFT abnormalities developed during feeding, specific causes such as sepsis or a reaction to drug administration were considered. The feeding regimen was discontinued only if the abnormalities progressed very rap-
PATIENTS AND METHODS
All adults receiving TPN in Oxford during the 5-year period between 1980 and 1985 were studied prospectively. Records were made of the clinical details of each patient, the feeding regimen(s) used, any episodes of sepsis, and the outcome of feeding. The diagnosis of sepsis was made on the basis of fever, raised white cell count, known or suspected source of infection, and in many cases a positive microbiologic culture. Patients with known, antecedent chronic liver disease were excluded from the analysis.
Reprint requests: Miss P. J. Clarke, Nuffield Department of Surgery, John Radcliffe Hospital, Oxford OX3 9DU, UK. 54
Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on March 30, 2015
55
idly and discontinuation
was considered clinicallv showed a gradual but progressive increase in alkaline analyzed for the group as a phosphatase and AST levels and a small increase in whole. In addition, the data were analyzed for a number plasma bilirubin, and after 4 weeks of feeding the biliof subgroups and a comparison made between the follow- rubin was abnormal in 31%, the AST in 27%. and the ing : (1) patients whose pre-TPN LFTs were normal and alkaline phosphatase in 32%. those whose results were abnormal; (2) patients who had episodes of sepsis during the course of parenteral feeding Effect of Episodes of Sepsis During TPN and those who did not; (3) patients who survived to leave the study, 208 patients (49.5%) had an episode hospital and those who died during, or soon after TPN of During LFTs in patients without sepsis are shown in sepsis. stopped; and (4) patients who were fed for more than 28 3 and show a similar pattern to that in the total Figure days were considered as a separate group. Half of all the patients without sepsis Finally, an additional analysis was performed on the group (Fig. 1). abnormal LFTs. The percentages eventually developed 308 patients with normal pre-TPN LFTs, who received of septic and nonseptic patients with abnormal LFTs at A TPN during the last 3 years of the study. detailed each week are shown in Figure 4. The trend was for study was made of the 64 patients in this group whose of LFTs LFTs became abnormal during feeding in order to ascer- septic patients to have greater derangements but the only statistically signiftain whether there was any common factor which pre- than nonseptic patients, icant difference between the two was a higher disposed to this problem. Statistical analyses of the data bilirubin level for weeks 2 and 3 ingroups septic patients (p were performed using the Mann-Whitney U-test x2 anal4
